HomeSmart Grid T&DEnergy EfficiencyStorage systems for utility-scale solar and wind

Storage systems for utility-scale solar and wind

Experts from Mott MacDonald offer lessons learnt in mitigating real-time power output fluctuation for solar PV and wind power plants using battery storage systems, and spotlight a storage project in Japan

Mount Komekura solar photovoltaic power plant in Japan

Credit: TEPCO

In some countries where peak demand occurs during the night, an energy storage system can be used to store energy generation from solar PV plants during the day and inject stored energy to the distribution system at night.

The energy storage system (ESS) also can be used for ancillary services such as voltage support, frequency control and load smoothing, as well as ramp rate control in order to maintain grid stability. Controlling power output fluctuation from renewable power generation can lower disturbances to the grid and decrease feeder tripping, which consequently decreases revenue loss.

Although the ESS benefits the grid system in terms of the intermittency of the power supply, plus provides greater control of the network, integrating the ESS significantly increases the total lifetime – capital, operation and maintenance – cost of projects. Additionally, an ESS typically has round-trip efficiencies of between 70 and 95 per cent, so will discharge less energy back into the system than it can absorb.

Limited utility-scale experience of ESS in renewable energy applications also raises uncertainties and risks, such as system failure causing loss of revenue for power plants.

Challenges for developers

One challenge for renewable power project developers relates to the efficient design of battery capacity to cost-effectively meet the grid utility’s requirements. To accurately size the ESS, it is recommended for the developer to base battery sizing on high-resolution data (i.e., one minute or less) of renewable resource or power output at the site location.

For utility-scale projects, it is industry practice for wind projects to install a meteorological mast at the site location prior to construction. With an appropriate setting, high-resolution data can be extracted at site and utilized for battery sizing. In contrast, it is currently not general practice for solar PV plants to have an on-site weather station before the plant enters operation. Even for countries with abundant national meteorological stations, irradiation data from such public data sources is mostly available at only an hourly, daily or monthly resolution.

Figure 1: Comparison of power output between a solar PV plant with and without ESS

Some of the major considerations in the design of an ESS are:

ࢀ¢ Required through cycle efficiency (this varies by technology);

ࢀ¢ Battery cell lifetime (which will vary by technology, use and installation condition);

ࢀ¢ The design and operation of the battery management system; and

ࢀ¢ The design and performance requirements of the power converter systems.

The ESS can be integrated into renewable power plants in multiple configurations depending on the type of technology used. Since the concept of this type of battery application is in the early stage of deployment, technical modelling is likely to have high uncertainty, which impacts on financial sensitivity scenarios and return on investment.

Based on Mott MacDonald’s experience, to date no energy management system (EMS) supplier provides any form of guarantee on annual curtailment loss. Related challenges include uncertainty in dynamic conditions and future data forecasting approach, as well as the limitation of the EMS algorithm to handle unpredicted changes. Such circumstances certainly pose technical risks to project owners and developers.

Challenges for project lenders

Efficiency, reliability and lifetime of the current battery storage systems over a typical power project life of 20 to 25 years have not yet been proven. Therefore, a technology review or bankability study of the ESS should be undertaken to define risks to project investors and lenders.

Such a study would need to review evidence regarding technology maturity, battery performance and degradation, system lifetime and warranty package from the system suppliers. As an example, let’s consider a detailed analysis of an ESS in Hokkaido, Japan. Hokkaido is an island with weak electrical interconnection to the main Japanese grid and dynamic weather conditions that can significantly vary the power output from solar photovoltaic (PV) plants.

Given higher geographical differences in wind resource patterns, the consequent size of battery storage systems in wind power plants can be extremely different depending on specific location, so this case study will focus on the analysis of battery sizing for solar PV plants. It should be noted, however, that the battery optimization concepts are applicable for both wind and solar power plants.

There were several technical assumptions for the Hokkaido project:

ࢀ¢ High efficiency lithium-ion battery;

ࢀ¢ Battery is connected to the AC system of the generation plant following conversion;

ࢀ¢ Battery is equipped with an independent inverter-transformer package;

ࢀ¢ Changes in plant power output limited at 1 per cent of plant AC capacity per minute;

ࢀ¢ Battery output always reaches battery rated power regardless of the state of charge and the charging/discharging duration.

There were also several financial assumptions and these are based on Mott MacDonald’s Asian and international project experience on solar PV and wind power plants with battery storage systems.

With the assumptions above, the power plant with the ESS gives a smooth increase and decrease in energy transmitted to the grid. In contrast, the plant without the ESS shows high fluctuation in power output. The purpose of battery storage in the system is for smoothing power output throughout the day to achieve compliance in the rate of change in power output stipulated by the grid utility.

Battery sizing technical analysis

If the rate of change of power output from the renewable power generation project exceeds the regulated level, the plant will be required to temporarily shut down or financial penalties will be imposed. Such opportunity loss, referred to as curtailment loss, will occur when the battery size – both battery capacity (MW) and time capacity (h) – is insufficient to absorb high resource fluctuations in some circumstances.

Mott MacDonald provides ranges of curtailment loss based on operating power plants. The simulation of the curtailment loss is based on four different battery rated capacity (MW) cases – 100 per cent, 80 per cent, 60 per cent and 40 per cent of power plant rated output.

Percentages of curtailment loss over total power output from different battery capacities with the maximum discharge time variation from 0.4 h to 0.6 h and 0.8 h are shown in Table 2.

It is obvious that high battery rated capacity and maximum discharge time can have significantly lower curtailment loss under the scenario considered. For various maximum discharge times at 100 per cent battery rated capacity, curtailment loss is around 0 per cent to 9 per cent of the project annual output. Curtailment loss increases to between 12 per cent and 36 per cent when battery rated capacity is reduced to 40 per cent of plant capacity, with maximum discharge time of 0.4h to 0.8h.

Figure 2. Levelized ESS cost for a solar PV plant against various ESS package cost scenarios

Battery sizing financial analysis

The combination of battery system cost and curtailment loss per unit installed battery capacity (MWh) reveal the representative expenditures for the battery system over the project lifetime. Considering net present value terms, a levelized ESS cost is used to justify the most suitable battery capacity option for a project.

The levelized ESS cost for the purpose of this article is expressed as the above. The example of levelized cost of each battery size for a solar PV plant is summarized in Table 3.

Even though the larger battery size results in lower curtailment loss, the investment and maintenance cost portions can outweigh the saving opportunity from curtailment loss. For this article, the battery with 80 to 100 per cent capacity and 0.4 h maximum discharge time appears to be the most cost-effective battery size for the project.

Due to differences in national grid regulations and requirements, as well as a wide range of project-specific parameters, Mott MacDonald strongly recommends this kind of levelized cost study to be explored project-by-project.

This article provides an example of the change in levelized ESS cost for a solar PV plant against various ESS cost scenarios as illustrated in Figure 2.

From this exercise, the levelized ESS cost range for a solar PV project decreases consistently along with the reduction in ESS package cost.

This article outlines a case study of ESS use to meet utility requirements, limiting the maximum ramp rate in change of power output from an intermittent renewable generator, to give an illustration of the parameters, costs and challenges involved.

To ensure optimization of the ESS wherever such utility requirements apply, Mott MacDonald recommends a levelized cost study to be performed project-by-project.

The lack of fine-resolution meteorological and operating data limits analysis and simulation, resulting in significant performance uncertainty. Additionally, a greater understanding of the practical ESS control algorithm performance when equipped with utility-scale renewable energy generation is still required.

Since application of utility-scale ESS occurs within a multidisciplinary framework, strong collaboration between the grid utility, developers, ESS manufacturers and all relevant parties is necessary for efficient infrastructure spending. For example, the requirement for ESS storage and ramp rate regulation can be reduced in cases where the grid utility benefits from accurate weather forecasting, either from the specific plant or for the broader renewable portfolio on the grid system. It could therefore predict production to adjust generation dispatch accordingly.

Natthida Jongsuwanwattana, Nonthi Cherdsanguan and Vipant Chayavichitsilp are all graduate renewable energy engineers at Mott MacDonald.